Cliff Wall Redux 2015

[Tangled-Universe]

Whats interesting to me is how the verticals are all actually vertical. They do not converge and cross anywhere. The shapes are clearly rectangular. But they are very narrow/thin. If they all could be wider to varying degrees in the horizontal, than would this not be a fully successful rectangular noise image? And can the horizontal be widened?

The noise functions are flattened. Unflattened they would simply look like where they came from, so either ridged or billowy or perlin-like. Unfortunately.

I have had succes with clamping the noise to get "blocks" and it works *perfect* on a flat planet/plane object.
However, any moment it has to follow a non-flat surface it becomes the story mentioned and discussed over and over again in the square noise threads; TG's texture system is 'projecting the noise top-down' over the terrain.
In Cry-engine you can use "tangent space normal mapping" which means that you can create a vdisp map/function which always displaces irrespective of its underlying surface.
We would need something similar in TG or at least a more "true 3D" way of surface interpretation by displacement shaders.

[TheBadger]

I see. I actually understood you too. Thanks for the clarity Martin.

Well, thats too bad though. I thought here and in ulco's thread, that there may be some testing of alpha stuff going on. But your post shows that people are just finding ways to force the current version to do more... Thats cool too, for sure, but I was really getting excited about a total solution to the original problem.

anyway, its still cool to watch the problem solving efforts, thats half the fun for me to come here. So thanks for that guys.